FSRS, the algorithm that decides when to review

What is FSRS?

FSRS stands for Free Spaced Repetition Scheduler: a free and open-source scheduler for spaced repetition. In practice, it is an algorithm that looks at your answer history and predicts, for each card, the date you are about to forget it — then schedules the review just before that point.

It is now the reference in the most demanding memorisation communities, and it is built into the main flashcard tools, Anki included. This is not an in-house recipe: it is a collaborative project, documented and open to scrutiny.

How it works: three variables

FSRS is built on a model called DSR, which describes the state of each card in your memory with three quantities:

• Difficulty — how hard this card is for you. A card you keep getting wrong is treated as harder.
• Stability — how long the memory holds before it fades. The more you get it right, the higher the stability and the longer the intervals.
• Retrievability — the probability that you can recall the answer at a given moment. FSRS triggers the review when that probability drops towards the threshold you are aiming for (often 90%).

Every answer you give updates those three values, and the next review date is recalculated. Nothing to tune by hand: the algorithm learns from your own results.

FSRS vs SM-2: what is the difference?

For decades, most apps used SM-2, an algorithm created by researcher Piotr Woźniak in 1987. SM-2 works, but it applies fixed rules and a single "ease factor" per card, the same for everyone.

FSRS goes further: it tunes its parameters to your way of remembering, from your actual answers. The result measured by community studies: for the same level of retention, you typically get 20 to 30% fewer reviews than with SM-2. In other words, less time spent on what you already know.

Why this matters to you

• You review neither too early (wasted time) nor too late (already forgotten): each reminder lands at the right moment.
• The schedule adapts card by card: easy items come back rarely, hard ones more often.
• You get the best retention for the least effort — the whole point of spaced repetition.

FSRS in Memnos

Memnos uses FSRS by default, with no configuration. You paste your notes or generate your cards, you answer honestly, and Memnos schedules every reminder for you. No ease factors to adjust, no obscure settings: the algorithm works in the background.

Frequently asked questions

What is FSRS?

FSRS is an open-source algorithm that works out the best moment to review each card, based on your own answers. More accurate than the older SM-2, it clearly reduces the number of reviews needed for the same retention. In Memnos it is on by default, with nothing to configure.

Which apps use FSRS?

Anki has it built in, and Memnos uses it by default with no configuration. Since FSRS is open-source, other flashcard tools have adopted it since.

Is FSRS complicated to use?

No. The complexity is in the algorithm, not in the using. You answer your cards, the algorithm does the rest. In Memnos there is nothing to set up.

Does FSRS replace SM-2?

It is its modern successor. SM-2 is still decent, but FSRS is more accurate because it adapts to your memory instead of applying fixed rules.

Do you need a lot of data before FSRS is useful?

It works from day one with solid default values, and gets more accurate the more you review.